Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature Request: profile output should be sorted #7638

Closed
steven-johnson opened this issue Jun 15, 2023 · 2 comments
Closed

Feature Request: profile output should be sorted #7638

steven-johnson opened this issue Jun 15, 2023 · 2 comments

Comments

@steven-johnson
Copy link
Contributor

It would be a bit easier to read the profile output if halide_profiler_report_unlocked() sorted the output by percentage.

@abadams
Copy link
Member

abadams commented Jun 15, 2023

The downside of that is it would make the ordering unstable from run to run, making comparisons difficult. Currently it's in rough order of realization, which is also annoying, because it's sensitive to the schedule.

Possibly it should be in pipeline order, so early stages are always at the top, and later stages are always at the bottom.

@steven-johnson
Copy link
Contributor Author

would make the ordering unstable from run to run

True, but usually when I'm profiling I want to start by looking at the biggest costs; this makes them easier to spot. (Maybe consider adding it via an env var?)

in pipeline order

Is that information transmitted to the profiler right now? if not, how hard would it be?

ardier pushed a commit to ardier/Halide-mutation that referenced this issue Mar 3, 2024
Fixes halide#7638) (halide#7639)

* Allow optional sorting of profiler output via HL_PROFILER_SORT env var (Fixes halide#7638)

* trigger buildbots

* Update profiler_common.cpp

* Update float16_t.cpp

* Update float16_t.cpp

* Update float16_t.cpp

* Update float16_t.cpp
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants